Unique Roles for E2F1 in the Mouse Lens in the Absence of Functional pRB Proteins

Abstract

purpose. Normal lens fiber cell differentiation requires functional retinoblastoma protein (pRB), because inactivation of this protein results in proliferation and apoptosis in normally postmitotic, differentiating fiber cells. Loss of either E2F1 or -3 can partially rescue the lens phenotype in Rb-deficient mice, implying that these E2Fs may have specific targets in this system. The purpose of this study was to determine what unique role E2F1 may play.

methods. Expression of E2F family members and target genes was analyzed in the lenses of nontransgenic, E2F1-null, αAE7;E2F1-sufficient; and αAE7;E2F1-null mice by in situ hybridization, Northern blot analysis, and RT-PCR.

results. In lenses of E2F1-null mice, there was no change in the expression of E2F-2 to -5 or their target genes, compared with E2F1-sufficient mice. However, in the lens of αAE7 mice where pRB proteins are inactivated, expression of E2F2 and -3a was increased. The E2F3a increase, but not that of E2F2, was dependent on E2F1. Expression of E2F target genes was increased with expression of E7 and expression of one of these, p19ARF, was E2F1 dependent.

conclusions. Although in the normal lens there do not appear to be unique roles for E2F1 that cannot be fulfilled by other E2F family members, in the absence of functional pRB proteins, E2F1 is specifically responsible for the increased expression of E2F3a and p19ARF. These findings suggest that E2F1 may be the preferred E2F regulating these target genes in the normal lens.

Spatial and temporal control of cellular quiescence, proliferation, differentiation, and the transition among these states is essential for proper embryonic development. The developing mouse lens is an ideal system for studying cell cycle control in vivo, because the lens contains defined regions composed of cells in specific stages of the cell cycle. The lens is composed of a layer of undifferentiated cells, the epithelium, which borders the anterior surface of a large mass of differentiated fiber cells. The cells in the central region of the epithelium are quiescent but are capable of entering the cell cycle. More peripheral cells in the germinative zone actively proliferate and as these cells divide, they move posteriorly into the transition zone, where they withdraw from the cell cycle and differentiate into fiber cells. ^{1 2} Recent studies have identified the product of the retinoblastoma susceptibility gene (pRB) as a regulator of lens fiber cell differentiation and have identified members of the E2F transcription factor family as potential targets for pRB action. ^{3 4 5} In this study, we examined the possible role of a known target of the pRB protein, the E2F1 transcription factor, in the lens in the presence or absence of functional pRB. 

Many studies have shown that the tumor-suppressor protein pRB is required for cell cycle withdrawal and differentiation of lens epithelial cells into fiber cells. Mutational inactivation of Rb leads to a failure of cells to withdraw from the cell cycle and differentiate; ultimately, they undergo apoptosis. ³ Similarly, expression of the E7 oncoprotein from human papillomavirus (HPV)-16, which binds to and inactivates pRB proteins in lens cells at the time they normally differentiate, using the mouse αA-crystallin promoter, results in a similar phenotype. ⁵ Expression in the fiber cells of the truncated form of the simian virus (SV)40 T antigen, which also binds pRB but not p53, leads to similar consequences. ⁴ Collectively, these studies demonstrate the importance of pRB and possibly pRB family members in regulating cell cycle withdrawal during fiber cell differentiation. 

pRB and the other family members, p107 and p130, when in their hypophosphorylated, active state, are capable of binding members of the E2F family of transcription factors. The pRB-E2F complex (along with the E2F binding partner, DP) ⁶ acts as a transcriptional repressor of some target genes, such B-myb and cdc2. ^{7 8} As the cell progresses through G1, pRB becomes hyperphosphorylated and dissociates from E2F, allowing the E2F-DP dimer to act as a transcriptional activator. E2Fs have been shown to activate many genes essential for progression into the S phase, as well as a variety of other cell-cycle control factors, including cyclins and other E2F family members. ⁶  

The E2F family of transcription factors is made up of six members that are defined by their similar DNA-binding domains. E2F1 to -5 also contain a transactivation domain and a pocket protein binding domain that allow them to interact with pRB, p107, or p130 ⁶ ; only E2F6 does not contain these domains. ^{9 10} All E2Fs share similar dimerization domains that allow them to heterodimerize with either DP-1 or DP-2, which is thought to be necessary for efficient DNA binding. ⁶ Although all E2F-DP dimers recognize similar DNA sequences, evidence suggests that not all E2Fs are capable of regulating all target genes. Specifically, E2F1, -2, and -3 activate target genes that are capable of inducing proliferation, whereas E2F4 and -5 do not possess this activity. ¹¹  

Given the known role of pRB and pRB family members as negative regulators of E2Fs’ activity and the essential role of pRB in lens fiber cell differentiation, it is likely that E2Fs represent important targets of pRB regulation in lens cell growth and differentiation. In the lens, E2F1 to -5 are expressed in the epithelium, whereas only E2F1, -3, and -5 are expressed in fiber cells. ¹² Therefore, E2F1, -3, and -5 may play a role in fiber cell differentiation. However, the absence of reported phenotypes in germline mutations in E2F1, -3, or -5 ^{13 14} suggest that these factors are dispensable for normal lens differentiation or that the absence of any one E2F in the fiber cells can be compensated for adequately by the other family members. Moreover, when pRB is inactivated, either through germline mutation or by expression of E7, the resultant proliferation and apoptosis in the fiber cell compartment can be partially rescued by either an E2F1-null mutation ^{15 16 17} or an E2F3-null mutation ¹⁸ —supporting evidence for at least some roles for E2F1 and -3 that cannot be completely fulfilled by other E2F family members. Together, this evidence suggests that in the normal lens, where family members probably compensate for the loss of any particular member, it is difficult to separate the roles of individual E2Fs, whereas in the situation in which pRB function is disrupted and E2F activities are presumably increased, the unique roles of the individual E2Fs may be apparent. 

In this study, we sought to determine whether compensation for the loss of E2F1 could be accounted for by increased expression of other E2Fs. In addition, we sought to determine whether any unique role for E2F1 could be discerned in the absence of functional pRB protein. To this end, we used the αAE7 mouse as a model for pRB inactivation in the lens, because germline mutation of Rb leads to embryonic lethality early in lens development. Through a series of experiments to examine E2F and E2F target gene RNA levels in the lenses from control, E2F1-null, αAE7, and αAE7;E2F1-null mice, we show that loss of E2F1 alone did not significantly alter the levels of expression in the lens of either other E2Fs or that of their target genes. In our study, expression of E7 led to increased expression of E2F2 and -3a, as well as increased expression of all target genes examined. Finally, although the E7-induced increases in expression of most of the genes examined was only partially E2F1 dependent, the E7-induced increases in expression of p19ARF and E2F3a were entirely E2F1 dependent. Thus, E2F1 mediates effects resulting from the loss of pocket protein function by uniquely regulating the expression of a subset of E2F target genes, suggesting that there may be E2F1-dependent targets in the normal lens. 

Materials and Methods

Mice

Mice carrying the E2F1-null mutation and the transgenic mice expressing HPV-16 E7 specifically in the lens on both the E2F1-sufficient and E2F1-null background have been described previously. ^{5 13 16} Mice were genotyped for E7 and E2F1 status by PCR analysis of DNA prepared from tail biopsy specimens, as described. ⁵ All procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Reverse Transcription-Polymerase Chain Reaction

RNA was prepared from lenses dissected from a liter of neonatal mice, by using extraction reagent (TRIzol; Life Technologies, GibcoBRL, Grand Island, NY) according to the manufacturer’s directions, and stored at −80°C. RNA was quantified spectrophotometrically. RT-PCR was performed with first-strand synthesis beads (Ready-to-Go You Prime First Strand Synthesis Beads; Amersham Pharmacia Biotech, Piscataway, NJ) with the following primers: E2F2, as previously described ¹² ; E2F3a: 5′-CCC GCC CTG GAG CAG TA-3′; 5′-CCC AGT TCC AGC CTT CG-3′; E2F3b: 5′-GGC TGC TTT CGG AAA TGC-3′, 5′-TTT CAC AAC TAT AAC AGT TT-3′; and γC crystallin, as previously described. ¹⁹ Reaction products were electrophoresed in a 1% agarose gel in 1× TAE buffer (.04 M Tris, 1 mM EDTA, adjusted to pH 7.5–7.8 with acetic acid), transferred to nylon membranes (GeneScreen; NEN Life Science Products, Inc., Boston, MA), and hybridized. 

Northern and Southern Blot Analysis

RNA was prepared as for RT-PCR. Polyadenylated RNA was prepared from 20 μg total lens RNA from neonatal mice of the indicated genotypes, using a kit (Oligotex mRNA Mini Kit; Qiagen, Hilden, Germany), electrophoresed in a 1% agarose, 0.66-M formaldehyde gel in 1× 3-(N-morpholino)propanesulfonic acid (MOPS), and transferred to nylon membranes (GeneScreen Plus; NEN Life Science Products, Inc.) according to the manufacturer’s instructions. Hybridization and washes were performed according to the instructions. Random primed labeled probes were made with a kit (Random Primed DNA Labeling Kit; Roche Molecular Biochemicals, Mannheim, Germany) from restriction-digested fragments purified with a DNA purification system (Prep-a-Gene; Bio-Rad, Richmond, CA). End-labeled probes were made with T4 polynucleotide kinase (Promega, Madison, WI). Expression levels were quantitated by a phosphorescence imager (PhosphorImager; Molecular Dynamics, Sunnyvale, CA) and analyzed on computer (ImageQuant; Molecular Dynamics, and Excel; Microsoft, Redmond, WA). 

Generation of Probes

To generate probes, the following primer sets were used to amplify the indicated gene from mouse genomic DNA and were cloned into a vector (P-gemT; Promega) Plasmids were then sequenced. E2F2 and E2F5: as has been described ¹² ; E2F3a: 5′-ATG AGA AAG GGA ATC CAG C-3′; 5′-TCC TGG TGC TGG TGG CTG-3′; E2F3b: 5′-CTC CCC CGG AGC CAG GCT GCT TTC GGA AAT GCC CTT ACA GCA GCA GCA G-3′, p19ARF: 5′-GAG TAC AGC AGC GGG AGC ATG G-3′; 5′-GGA TTC CGG TGC GGC CCT CTT-3′. The E2F4 probe was derived from plasmid mE2F4-3′ (provided by Rene Bernards, The Netherlands Cancer Institute, Amsterdam, The Netherlands). Probes for hGAPDH, hCyclin A2, B-myb, and cdc-2 were derived from plasmids pBSGAPDH, hCylA, mB-myb, and mCdc2 (provided by Peggy J. Farnham, University of Wisconsin, Madison, WI). The hCylA AvaII fragment was subcloned into the EcoRV site of pBS to isolate regions with high homology to the mouse sequence. The probe for N-myc was derived from the plasmid pMN1.1 (provided by Ronald A. DePinho, Dana Farber Cancer Institute, Boston, MA). The probe for p53 was derived from plasmid p53 3/2, as previously described. ¹⁹  

In Situ Hybridization

In situ hybridizations were performed as described previously. ²¹ E2F2 sense and antisense probes were synthesized using T7 or SP6, respectively, from the plasmid just described, with a kit (MAXIscript; Ambion, Austin, TX) to generate [α-³⁵S] uridine triphosphate (UTP)-labeled riboprobes. DP-1 sense and antisense probes were similarly synthesized with T7 and T3, respectively, from the plasmid pBSmDP-1 (provided by Peggy J. Farnham). N-myc sense and antisense probes were similarly synthesized using T3 and T7, respectively, from the described plasmid. Hybridized sections were exposed to emulsion (NTB-2; Eastman Kodak, Rochester, NY) in the dark for 2 to 3 weeks before developing. After they were developed, the sections were counterstained with hematoxylin, mounted, and viewed under both light-field and dark-field illumination. 

Results

Expression of E2F3 Isoforms in the Mouse Lens

By RT-PCR using primers specific to the C-terminal half of E2F3, Rampalli et al. ¹² have shown that E2F3 is expressed in the epithelium and the fiber cell compartment of the lens. Recent studies have shown that there are two isoforms of E2F3 referred to as E2F3a and -3b. ²² E2F3b differs from E2F3a only in the use of an alternate first exon. Therefore, the previous work would not have distinguished between the two isoforms. To determine which isoforms of E2F3 are expressed in the lens, we designed 5′ primers to the different first exons that would specifically amplify either E2F3a or -3b. RT-PCR then was performed with total RNA isolated from either epithelial cells or fiber cells. The reaction products were subsequently hybridized with probes for either exon 1a or exon 1b, to confirm the identity of the amplified product. Figure 1A shows the relative locations of the primers and probes used. Fragments of the expected sizes for E2F3a and -3b were seen in both epithelial and fiber cell RNA (Fig. 1B) . To determine whether epithelium and fiber cells had been separated cleanly, E2F2 and γC crystallin were also amplified by RT-PCR and hybridized with specific probes by Southern blot. As expected, E2F2 transcripts were found exclusively in the epithelium, whereas γC crystallin transcripts were found only in the fiber cells (Fig. 1C) . Based on these data, we conclude that both E2F3 isoforms are expressed in the epithelial and fiber cell compartments of the mouse lens. 

Expression of E2F Family Members in the E2F1-Null Lens

It has been shown that E2F1 is expressed in both the epithelium and fiber cells of the mouse lens, ¹² yet the E2F1-null mouse appears to have no lens defect. ¹³ A possible explanation for this finding is that other E2F family members are able to compensate for the loss of E2F1, potentially through upregulation of their expression levels. To examine this possibility, Northern blot analysis was performed for other E2F family members using lens RNA from either E2F1-sufficient or E2F1-null mice. Bands of the expected sizes for E2F3 to -5 were seen in RNA from both nontransgenic and E2F1-null lenses (Fig. 2) . There was no significant change in expression of E2F3 to -5 with the loss of E2F1. E2F2 was not detected in either the nontransgenic or E2F1-null lens by Northern blot or in situ hybridization (Fig. 3) , although E2F2 was detected by RT-PCR in RNA from nontransgenic mice (Fig. 1C) , indicating that E2F2 is expressed in the lens as previously reported, ¹² but at levels below the sensitivity of either Northern blot analysis or in situ hybridization. From these data, we conclude that expression levels of other E2F family members are not increased with the loss of E2F1. 

Effect of E7 on E2F Expression in the Presence and Absence of E2F1

It has been demonstrated that expression of the HPV-16 E7 oncoprotein in the lens results in proliferation and apoptosis in spatially inappropriate regions of the lens, and that this activity is dependent on E7’s ability to bind and inactivate pRB and pRB family members. ⁵ E7 expression also results in an increase in E2F1 expression throughout the fiber cell compartment. ²¹ Therefore, we asked if the expression levels of other E2F family members were similarly affected by E7 expression. Northern blot analysis with probes specific for each E2F (Fig. 2) indicates that expression levels of E2F2 and E2F3a were increased in the lenses of αAE7 mice compared with lenses of nontransgenic mice. In situ hybridization for E2F2 confirmed these results (Fig. 3) . The expression levels of E2F3b, -4, and -5 were not affected by expression of E7. 

It has been shown that loss of E2F1 provides a partial rescue of the proliferative and apoptotic defects in lenses of αAE7 mice. ¹⁶ However, significant levels of inappropriate proliferation and apoptosis remain in the lenses of αAE7;E2F1-null mice. Possible mediators of the remaining proliferation and apoptosis are the other E2F family members, particularly E2F2 and -3a, because they show increased expression in the αAE7 lens and are known to be capable of inducing proliferation. Both E2F2 and -3a have E2F-binding sites in their promoters. ^{22 23} Therefore, it is conceivable that their deregulated expression in the lens of αAE7 mice is dependent on E2F1. To explore this possibility, we compared the expression of the E2F family members in lenses of αAE7;E2F1-sufficient mice and αAE7;E2F1-null mice. Northern blot analysis with a probe specific for E2F3a showed that the increased expression level of E2F3a in the αAE7 lens was reversed by loss of E2F1 (Fig. 2) . In contrast, in situ hybridization for E2F2 demonstrated that E2F2 showed similar levels of increased expression in the lenses of αAE7 and αAE7;E2F1-null mice (Fig. 3) . As in the lenses of αAE7 mice, levels of E2F3b, -4 and -5 in the αAE7;E2F1-null lens were not significantly altered compared with their levels in lenses of nontransgenic mice (Fig. 2) . These data demonstrate that expression of only a subset of E2Fs shows increased expression in the presence of E7, with the expression of one of these, E2F3a, being entirely dependent on E2F1. 

Expression of the E2F-Binding Partner DP-1

The finding that expression of E7 in the lens leads to increased expression of some E2F family members, suggests that there is also increased E2F activity in the lenses of αAE7 mice. However, this model assumes that expression of E2Fs necessarily corresponds to E2F activity. The transcription-regulatory activity of the E2Fs depends on their ability to bind DNA. Heterodimerization with either DP-1 or DP-2 is required for E2Fs to bind DNA efficiently. ²⁴ Consequently, sufficient levels of a DP family member must be present in the same regions of the lens as the E2Fs for there to be any E2F activity. Therefore, we examined expression of DP-1 to determine whether there is the possibility of increased E2F activity in lenses expressing E7. We chose to examine the expression of DP-1, because it is capable of heterodimerizing with E2F1 to -5, unlike DP-2, which is not thought to heterodimerize with E2F5. 

We examined expression of DP-1 by in situ hybridization. In lenses from nontransgenic mice, DP-1 transcripts were found in the epithelium and the fiber cells of the transition zone. In the lenses of αAE7 mice, the expression pattern of DP-1 was expanded to include the entire fiber cell compartment (Fig. 4) . We also examined the expression of DP-1 in the lenses of E2F1-null mice and αAE7;E2F1-null mice. As expected, the expression pattern of DP-1 was not changed with the loss of E2F1. These data indicate that in lenses of both αAE7;E2F1-sufficient and αAE7;E2F1-null mice, there is DP-1 expression, suggesting that E2F activity could be found throughout the fiber cell compartment. 

Effect of Loss of E2F1 on the Expression of E2F Target Genes

To determine whether loss of E2F1 affects expression levels of E2F target genes, we examined the expression levels of known E2F target genes in nontransgenic and E2F1-null lenses. E2F target genes that are thought to be targets of E2F-mediated repression (cdc2, B-myb) ^{7 8} as well as activation (p19ARF, cyclin A2, N-myc) ^{25 26} were selected for Northern blot analysis. Because p53 is thought to act downstream of E2F1 in a proapoptotic pathway, ^{27 28} expression of p53 was measured as well. Lens RNA from E2F1-sufficient and E2F1-null mice was analyzed by Northern blot using probes specific for each of these genes (Fig. 5 ). N-myc expression was also analyzed by in situ hybridization (Fig. 6) . No significant difference in expression levels of the examined target genes was observed in the lenses of E2F1-null mice compared with the lenses of E2F1-sufficient mice. These data indicate that the expression of these target genes is not strictly dependent on the expression of E2F1 and that there are other activities in the lens that can compensate for both the transcriptional activating and repressive functions of E2F1. 

E2F Target Gene Expression in the αAE7 Lens in the Presence and Absence of E2F1

Because E7 expression leads to increased expression of E2F1, -2, and -3a, it is reasonable to assume that there is also an increase in E2F activity, which would lead to increased expression levels of E2F target genes. Therefore, to look for evidence of increased E2F activity, we compared the expression levels of E2F target genes in the lenses of αAE7 mice with the expression levels in lenses of nontransgenic mice. Expression levels of cdc2, B-myb, cyclin A2, p19ARF, N-myc, and p53 were assessed in lenses of nontransgenic and αAE7 mice by Northern blot analysis (Fig. 5) and quantified by phosphorescence imager. Expression levels of cdc2, B-myb, cyclin A2, p19ARF, and p53 were increased in the presence of E7. This result is consistent with the model of increased E2F activity in the αAE7 lens. The expression levels of N-myc were not significantly altered with the expression of E7, according to Northern blot analysis. However, in situ hybridization showed that the N-myc pattern was clearly expanded throughout the fiber cell compartment rather than restricted to the epithelium and transition zone (Fig. 6) . 

Because it has been shown that E2F1 can mediate some of the E7-induced proliferation and apoptosis, it is possible that expression of E2F targets are at least partially dependent on E2F1. To test this possibility, we compared expression levels of E2F target genes in lenses expressing E7 on both the E2F1-sufficient and deficient backgrounds by Northern blot analysis (Fig. 5) and quantified the results by phosphorescence imager (Fig. 7) . Expression levels of cdc2, B-myb, cyclin A2, and p53 were lower in the lenses of αAE7;E2F1-null mice than in the lenses of αAE7;E2F1-sufficient mice. However, the expression levels of these genes in lenses of αAE7;E2F1-null mice was still higher than their levels in lenses of nontransgenic mice. For p53, these results were verified by in situ hybridization (data not shown). The increased expression of p19ARF in lenses of αAE7 mice was completely reversed in lenses of αAE7;E2F1-null mice. By in situ hybridization, the expression pattern of N-myc in lenses of αAE7;E2F1-null mice was not altered, compared with the expression pattern in lenses of αAE7;E2F1-sufficient mice (Fig. 6) . Thus, the effect of E2F1 on E7-induced changes in E2F target gene expression differed among the target genes examined. For some target genes (cdc2, B-myb, cyclin A2, and p53) the increased expression levels were partially dependent on E2F1; for others the change in expression was either entirely independent of E2F1 (N-myc) or entirely dependent on E2F1 (p19ARF). 

Discussion

Previous studies have shown that inactivation of pRB function in the fiber cells of the lens leads to continued proliferation, failure to differentiate normally, and apoptosis. Increased activity of the E2Fs is thought to play a role, at least in mediating these effects. However, it is not known whether individual E2Fs play specific roles in this process, because mutational inactivation of either E2F1 or -3 partially rescues this effect. In this study there was no change in expression levels of either E2F family members or E2F target genes in the lens when E2F1 was mutationally inactivated. In contrast, E2Fs associated with proliferation and all E2F target genes examined showed increased expression in the absence of functional pRB proteins. Increased expression of most of the genes examined (E2F2, cdc2, B-myb, cyclin A, p53) showed only a partial dependence on E2F1. However, E7-induced expression of two genes, E2F3a and p19ARF, was completely dependent on E2F1. Thus, in the absence of functional pRB, E2F1 contributed to the regulation of expression of many E2F target genes. In addition, E2F1 played a unique role in regulating expression of certain E2F target genes in the absence of functional pRB. This suggests that these latter targets may be preferentially regulated by E2F1 normally. 

Expression of E2F3 Isoforms in the Lens

In this study, E2F3a and -3b were expressed in both the epithelium and fiber cells of the lens. It is interesting that E2F3a was expressed in both the fiber cells and the epithelium, because E2F1 was also expressed in both compartments. It is known that E2F3a has activities similar to those of E2F1. ¹¹ Therefore, E2F3a is a good candidate to provide compensatory E2F activity in the E2F1-null lens. Because E2F3b has only recently been described, little is known about its activity. However, given the knowledge that expression of E2F3b is not cell-cycle regulated, ²² it is tempting to speculate that it performs its major role in cells that are not actively cycling, which in the lens would include cells in the central epithelium, transition zone, and fiber cell compartments. Further work determining the spatial distribution of E2F3a and -3b, as well as their relative expression levels in the two compartments of the lens, may provide some additional clues. 

Compensation for the Loss of E2F1

E2F1, -3, and -5 are all expressed in the fiber cells in the developing rat lens. ¹² However, there are no reported lens defects in mice with germline mutations in any one of these genes. ^{13 14} The most likely explanation for this is that other E2F family members are capable of compensating for the loss of another family member. Compensation could occur at the level of gene expression or activity. Expression of E2F family members in the lens did not change with the loss of E2F1, nor did expression of the required binding partner DP-1. Together, these data imply that although other family members are not transcriptionally upregulated to compensate for the loss of E2F1, they are still expressed with a DP-binding partner and therefore could functionally compensate for the loss of E2F1. In fact, the expression level of E2F target genes was not changed in the lenses of E2F1 null mice, providing further indication that this is likely. 

Regulation of E2F2 and -3 Expression

Previously, we have shown that expression of E7 in the lens results in proliferation and apoptosis in the fiber cell compartment ⁵ and that the increased proliferation and apoptosis are partially dependent on E2F1. ¹⁶ Although it has been shown that deregulated expression of E2F2 in fiber cells can lead to a similar phenotype as that in lenses with deregulated expression of E2F1 ²⁹ or expression of E7, it was not understood whether deregulated expression of E2F2 contributes to the phenotype of lenses with compromised pRB function. In this study, RNA levels of both E2F2 and -3a were increased in the presence of E7 but only the E7-induced increase in E2F3a expression was entirely dependent on E2F1. Therefore, if other E2Fs mediate the remaining proliferation and apoptosis seen in the lenses of αAE7;E2F1-null mice, E2F2 is the most likely candidate. Alternatively, it is conceivable that in the absence of functional pRB, E2F3b or -5 may activate targets that would support proliferation. 

It is not surprising that E2F1, -2, and -3a are the only E2Fs to show an increase in expression in the αAE7 lens, because only these E2Fs have E2F binding sites in their promoters. E2F3b, -4, and -5 have no known E2F binding sites in their promoters, and therefore their expression should not be affected by changes in E2F expression or activity. Therefore, E7-induced changes in expression levels or activities of E2F1, -2, and -3a would not be expected to effect E2F3b, -4, or -5. ²² However, because only E2F2, and not E2F3a, shows an E2F1-independent response to expression of E7, the regulation of these two E2Fs must be different. Increased expression of E2F3a is dependent on E2F1, but normal expression of E2F3a is not, and therefore it must be regulated by factors whose activity is not affected by expression of E7. E2F3b is a possible candidate to regulate expression of basal levels of E2F3a, because E2F3b’s expression levels are unaffected by loss of E2F1 or expression of E7. In addition, E2F3b contains the same DNA-binding domain as E2F3a and thus should be able to bind the E2F-binding sites in the promoter of E2F3a. However, it is also possible that E2F3a’s expression is normally regulated by a non-E2F factor. 

Unique Roles for Individual E2Fs in the Lens Fiber Cells

Previous studies have shown that E2Fs are responsible, at least in part, for mediating the effects of inactivated pRB. ^{16 17} However, it is not possible from these studies to discern distinct roles for individual E2Fs. By determining the levels of RNA expression for E2Fs and a battery of E2F target genes in pRB-compromised lenses in the presence and absence of E2F1, we have uncovered potential unique roles for E2F1. Based on their expression levels in the lenses of αAE7;E2F1-null mice compared with that in the lenses of αAE7;E2F1-sufficient mice, we grouped E2F target genes into three categories: those whose expression is partially dependent on E2F1, those that require E2F1, and those that are completely E2F1 independent. Most of the target genes examined fall into the first category. Cdc2, B-myb, and cyclin A2 all showed increased expression in the αAE7;E2F1-null lens, but at a level slightly lower than that in the αAE7;E2F1-sufficient lens. This indicates that for these genes, E2F1 plays a role in the E7-induced increased expression, but that other factors, perhaps E2F2, also contribute to their deregulated expression. It is of note that the degree of E2F1 dependency among the examined target genes was not the same (Fig. 7) , suggesting that the decrease in E7-induced expression results from the loss of specific transcriptional activities and is not a secondary effect resulting from the decreased proliferation in the αAE7;E2F1-null lens. 

Target genes in the second category, those that require E2F1, are E2F3a, as discussed earlier, and p19ARF. p19ARF is not detected in the lenses of either nontransgenic or E2F1-null mice, but shows a marked increase in expression in the lenses of αAE7 mice. This increase is completely dependent on E2F1, because p19ARF is not detected in the lenses of αAE7;E2F1-null mice. This is consistent with previous reports that p19ARF is a unique target of E2F1 and not the other family members. ²⁵ p19ARF has been shown to stabilize p53 protein levels, because of its ability to block mdm2-mediated p53 degradation. Therefore, it is proposed that p53 acts downstream of p19ARF in a proapoptotic, tumor-suppressor pathway. ^{27 28} According to this model, increased activity of p53 when p19ARF is expressed could contribute to p53-dependent apoptosis. Based on our results, we conclude that the loss of p19ARF expression and the consequent decrease in p53 stability leads to a quantitative reduction in p53-dependent apoptosis in the lenses of αAE7;E2F1-null mice compared with that in αAE7 mice. This conclusion is consistent with previous studies showing that, although loss of E2F1 reduces p53-dependent (as well as p53-independent apoptosis) in the lens of αAE7 mice, this reduction is only partial. ¹⁶  

The expression level of p53 mRNA may also play a role in the partial rescue of E7-induced apoptosis in the αAE7;E2F1-null lens. Although p19ARF expression requires E2F1, increased expression of p53 is only partially dependent on E2F1. Thus, in the lens fiber cells of the αAE7 mice, there must be factors in addition to E2F1 that are capable of inducing expression of p53 and consequently p53-dependent apoptosis. It is possible that, in contrast to tissue culture systems, E2F family members other than E2F1 are capable of inducing apoptosis in vivo. Because E2F2 shows increased expression in the αAE7 and αAE7;E2F1-null lens, it is a potential mediator of this proapoptotic activity. In addition, because apoptosis in Rb-null lenses can be partially rescued by an E2F3-null mutation, ¹⁸ E2F3 must have proapoptotic role as well. 

The third category of E2F target genes is that in which expression is completely independent of E2F1. N-myc is a potential target for E2Fs, because E2F consensus binding sites are located within the N-myc promoter. ^{26 30} Northern blot analysis showed that there was no significant change in the level of N-myc RNA when E7 was expressed in the presence or absence of E2F1, implying that despite the presence of E2F binding sites in its promoter, N-myc is not a target of E2F1, -2, or -3a regulation in the lens. In situ hybridization showed clearly that N-myc had an expanded pattern of expression in the αAE7 lens that was not dependent on E2F1 status. It has been shown in the chick lens that N-myc expression increases during the early steps of withdrawal from the cell cycle by the fiber cells, but is repressed once these cells have fully elongated and differentiated. ³¹ Therefore, in the αAE7 lens where fiber cells do not withdraw from the cell cycle or differentiate, we suggest that there is neither a local increase in N-myc expression nor a repression of its expression. The result is an expanded expression pattern for N-myc without any significant changes in total N-myc RNA levels. 

In summary, our study shows that in the presence of functional pRB proteins, there is probably compensation for the loss of E2F1 at the level of E2F activity. In contrast, in the absence of functional pRB proteins, unique roles for E2F1 become apparent. Under this latter condition, expression of numerous E2F targets genes, including specific E2Fs themselves, were differentially regulated by E2F1. An important finding is that the changes in expression of two of the genes studied, E2F3a and p19ARF, in lenses where pRB function was compromised, appeared to rely entirely on E2F1. Also of importance, the finding that E2F2 expression was increased in the absence of functional pRB proteins in an E2F1-independent manner suggests another factor that may contribute to the proliferation and apoptosis observed in pRB-compromised fiber cells. The results of our study suggest that under normal conditions during lens fiber cell differentiation, multiple E2Fs may participate in regulation of some target genes, whereas others may be preferentially regulated by E2F1. Future studies are necessary to determine the normal in vivo nature of pRB-E2F-target gene regulation in lens fiber cell differentiation. 

 Supported by National Eye Institute Grant EY-09091 (AEG).

 Submitted for publication September 28, 2001; revised January 15, 2002; accepted January 18, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Expression of E2F3 isoforms in neonatal mouse lenses. (A) Diagram of E2F3a and -3b. Primers and probes used for RT-PCR and Southern blot analysis are indicated. (B) Total RNA (5 μg) from either fiber cells or epithelial cells (Epi) was used for RT-PCR, in the presence (+) or absence (−) of RNase, with primers specific for the indicated E2F3 isoforms. RT-PCR products were hybridized using the indicated probes. (C) RT-PCR for E2F2, which is expressed only in the epithelium, was performed using 1 μg and hybridized with a gene-specific probe, in the presence (+) or absence (−) of RNase. RT-PCR for γC crystallin (γC Cry), which has fiber-cell-specific expression, was performed with 1 ng RNA and hybridized with a gene-specific probe. All probes were [α-³²P] dCTP random prime-labeled probes specific for the indicated genes, except for E2F3b, for which a [γ-³²P] adenosine triphosphate (ATP) end-labeled probe was used.

Figure 2.

 

Northern blot analysis of E2F family members. Polyadenylated RNA isolated from 20 μg total RNA (A, C) or 10 μg total RNA (B) from nontransgenic (NT), E2F1-null (E2F1−/−), αAE7 (E7), and αAE7;E2F1-null (E7, E2F1−/−) neonatal lenses was electrophoresed, transferred to nitrocellulose membranes, and hybridized with [α-³²P] dCTP random prime-labeled probes specific for the indicated genes, except in the case of E2F3b, for which a [α-³²P] adenosine triphosphate (ATP) end-labeled probe was used. Expression of each gene was standardized to expression in the NT lens with a phosphorescence imager. GAPDH was used as an internal control.

Figure 3.

 

In situ hybridization analysis of expression of E2F2. Paraffin-embedded sections (5 μm) of neonatal eyes from nontransgenic (A–C), E2F1-null (D–F), αAE7 (G–I), and αAE7;E2F1-null (J–L) mice were hybridized with an [α-³⁵S] UTP-labeled E2F2-specific antisense (A, B, D, E, G, H, J, K) or sense (C, F, I, L) primer. Slides were exposed for 2 weeks, processed, and viewed with bright-field (A, D, G, J) and dark-field (B, C, E, F, H, I, K, L) microscopy. All lenses are oriented with the anterior at the top. e, epithelium; f, fiber cells; r, retina. Bar, 100 μm.

Figure 4.

 

In situ hybridization analysis of expression of DP-1. Paraffin-embedded sections (5 μm) of neonatal eyes from nontransgenic (A, B), E2F1-null (C, D), αAE7 (E, F), and αAE7;E2F1-null (G, H) mice were hybridized with an [α-³⁵S] UTP-labeled DP-1-specific antisense (A, C, E, G) or sense (B, D, F, H) primer. Slides were exposed for 2 weeks, processed, and viewed under dark-field microscopy. e, epithelium; f, fiber cells; r, retina. All lenses are oriented with the anterior at the top. Bar, 100 μm.

Figure 5.

 

Northern blot analysis of E2F target genes. Total RNA (10 μg) from nontransgenic (NT), E2F1-null (E2F1−/−), αAE7 (E7), and αAE7;E2F1-null (E7, E2F1−/−) neonatal lenses was electrophoresed, transferred to nitrocellulose, and hybridized with an [α-³²P] dCTP random prime-labeled probe specific for the indicated gene. Expression levels of the examined genes in RNA from each genotype were then compared with expression levels in the NT lens with a phosphorescence imager. GAPDH was used as an internal standard.

Figure 6.

 

In situ hybridization analysis of expression of N-myc. Paraffin-embedded sections (5 μm) of neonatal eyes from nontransgenic (A, B), E2F1-null (C, D), αAE7 (E, F), and αAE7;E2F1-null (G, H) mice were hybridized with an [α-³⁵S] UTP-labeled N-myc-specific antisense (A, C, E, G) and sense (B, D, F, H) primer. Slides were exposed for 2 weeks, processed, and viewed with dark-field microscopy. All are lenses oriented with the anterior at the top. e, epithelium; f, fiber cells; r, retina. Bar, 100 μm.

Figure 7.

 

Quantification of Northern blot analysis for E2F target genes. Northern blot analysis containing RNA from nontransgenic (NT), E2F1-null (E2F1−/−), αAE7 (E7), and αAE7;E2F1-null (E7;E2F1−/−) from neonatal mice were hybridized with probes for the indicated E2F target genes. Signal intensities were normalized to GAPDH. Normalized signal intensities were compared with that in the nontransgenic lens (set to 1). In the case of B-myb, no signal was detected in RNA from nontransgenic lenses, and signal intensity in the αAE7 RNA was therefore compared with that in αAE7;E2F1-null RNA (set to 1). *Significant difference (P < 0.05) compared with the nontransgenic control; #significant difference (P < 0.05) compared with αAE7 RNA. n = 2 for all genes.

The authors thank Grace Panganiban for the use of her digital camera and Jennifer Drew and Carrie Graveel for helpful suggestions on the manuscript.